Paul Saffo says that communication technologies historically take 30 years or more to find their true purpose. Just look at how the Internet today is different than it was back in 1988. I am beginning to think this idea applies also to new computing technologies like artificial intelligence (AI). We’re reading a lot lately about AI and I think 2018 is the year when AI becomes recognized for its much deeper purpose of asking questions, not just finding answers.
Some older readers may remember the AI bubble of the mid-1980s. Sand Hill Road venture capitalists invested (and lost) about $1 billion in AI startups that were generally touted as expert systems. An expert system attempted to computerize professional skills like reading mammograms or interpreting oil field seismic logs. Computers were cheaper than medical specialists or petroleum geologists, the startup founders reckoned, so replacing these professionals would not only save money, it would allow much broader application of their knowledge. Alas, it didn’t work for two reasons: 1) figuring-out how experts make decisions was way harder than the AI researchers expected, and; 2) even if you could fully explain the decision-making process it required a LOT more computing power than originally expected. Circa 1985 it probably was cheaper to hire a doctor than to run a program to replace one.
But now, approximately 30 years later, AI has come back to life. Part of this is simply Moore’s Law. One million dollars worth of 1985 computational power costs less than a buck today, making those software experts way cheaper to run. But wait, there’s more! A second reason for AI’s resurgence is the availability of huge online data sets. Over the past 30 years nearly all the information that formerly resided on paper was reduced to electrons making true machine learning possible. This can’t be over-emphasized. Anyone my age remembers when early search engines indexed thousands of web pages, not billions. Highly technical data generally wasn’t available online in any volume but now it is.
The third major reason for AI’s resurgence is today we don’t even try to pick some doctor’s brain to build an expert system, instead empirically deriving skills directly from the data. Taking this one step further, we are moving to a system where we don’t even start with a question, just the data, allowing cloud-based systems to find what’s learnable from the data. That’s allowing AI to come up with its own questions and it’s the emerging trend on which I am trying to focus this prediction.
AI, analytics, and Big Data are keywords here. If you want to know more about this, early last year I wrote a series of three columns totaling more than 10,000 words on the topic. You can find those here.
If AI is going to figure out questions that need asking it is probably a mistake for me to start listing them here. But I can give you an idea what such questions would be like.
Look at this old chart I got from my friend Frank Starmer. It shows U.S. death rates from infectious and non-infectious diseases from 1900-1996. During this period we basically defeated tuberculosis and polio through vaccinations. We took a big bite out of influenza and pneumonia the same way (that big chart spike is the Influenza pandemic of 1918 — the Spanish Flu). We developed water and sewage treatment systems to defeat cholera and similar diseases. My mother, growing-up in Arkansas in the 1920s and 30s knew people who had chronic American-Grown malaria and that’s gone, too. The slight infectious uptick toward the end of this chart, in case you are wondering, represents the HIV/AIDS crisis, which peaked in those years, as well as an aging population that was still susceptible to pneumonia.
My reason for showing this chart is not because we did such a great job of curing infectious diseases but because we did such a poor job of curing everything else.
Everybody dies, of course. But this chart also predicts life expectancy. Look at 1900 and you’ll see there were 1700 total disease deaths per 100,000 population. That’s a 1.7 percent annual mortality rate suggesting a life expectancy of 100/1.7=58.82 years for both men and women. The 1996 total disease rate looks to be about 900 per 100,000 or 0.9 percent suggesting a life expectancy of 100/0.9=111 years! We don’t actually live that long for many reasons including infant mortality, wars, accidents, occupational, environmental, and lifestyle diseases, but it shows what’s possible with essentially zero contribution from improvements in non-infectious diseases.
If we found the data to extend this chart much of the news would be good thanks to the Clean Water Act, Clean Air Act, and fewer people smoking — not especially from improvements in medical treatments. Yet the cost of medical treatment continues to rise. Hopefully much of this represents our current investment in the very work I’m advocating here.
We’re getting to the point where AI should begin to organically suggest approaches that will help us improve medical outcomes. The big low-buck solutions like immunizations have for the most part already happened. Future gains will have to come incrementally, often one gene at a time. But that’s just where AI should shine, mining medical gems from all that data in our smart phones, fitness trackers, and home DNA tests.
Thirty years into AI, it’s time to start seeing significant rewards from this approach in many fields, not just medicine.
Great article Bob, one of my favorite predictions yet and I think you are spot on. I look forward to AI asking the tough questions that you haven’t considered yet, such as:
.
1.) “How long is ‘soon'”? and
2.) “When are you going to update everyone regarding the Mineserver ‘jihad'”?
@Chris D – Don’t forget how many weeks (months) are there in a “few days”?
@Mineserver Is it more than 5? Because that’s where we’re currently at:
.
“Look for a spec update and a new shipping schedule in a couple more weeks”
.
– Bob, January 2, 2018
The above post is another fake from the mine server trolls in here using another posters user name
Wait, which one?
That’s… that’s not how life expectancy calculations are done, Bob.
Try running the same calculation on the 1996 “All Causes” value, compare your result to every other organization of experts trying to calculate life expectancy, then try again. Try Figure 2a on the very study you linked to!
Good Lord, I overlooked that you weren’t even looking at the right line on the graph. You were looking at the “All Causes” line. It’s not a chart of deaths-by-disease. It’s a chart of all deaths, divided into infectious diseases and non-infectious causes. Causes, not diseases.
i r smartz?
It seems to me that many non-disease causes of death can be reduced by improving medicine, particularly if you include paramedic treatment and mental illness prevention and treatment.
As someone who once worked for an actuarial company, writing computer programs for actuarial calculations, my reaction to Bob’s analysis of life expectancy was, “Aaaaaagh! No!”
Unfortunately, I think he knows as little about AI as he does about calculating life expectancy.
Asking new questions and making new deductions requires creativity, which AI does not have and is nowhere near having.
The 800 lb gorilla in the room is the abject state of social etiology. The so-called “social sciences” are so deep in BS it’s over their head to think that they’re even in it. Arguments over “correlation vs causation” have, in the past, been perpetuated by such tragedies as violating the 10th Amendment so nothing remotely resembling control experiments exist — the Laboratory of the States. Control experiments are the traditional way of teasing apart correlation from causation*.
Now that the gold standard of unsupervised learning is understood to be lossless compression (algorithmic information), it _is_ feasible to work out etiology from the data alone — but only if you do so with a comprehensive set of longitudinal metrics including, not just social metrics, but ecological and bioinformatic metrics. The data is there. The theory is well known and advancing. The overwhelming need is there — and cascades into every aspect of life including AI application.
So why isn’t the X-Prize foundation or Kaggle or someone with serious bucks setting up something like the Hutter Prize for Lossless Compression of Human Knowledge, with the “corpus” being, basically, a unification of just about all of the well-cleaned databases about society, ecology and biology?
http://prize.hutter1.net/
*We can ignore the fact that giving people a choice of social policies under which they live is vastly more humane than torturing them that they must influence Washington D.C. to express their “consent” to experimental treatments. I ought to know as my group got The Launch Services Purchase Act of 1990 to privatize launch services signed into law by Bush Sr. only to have NASA basically ignore it as though it were an immigration law. That required non-trivial personal sacrifices from taxpaying citizens who have lives to lead and simply cannot compete with bureaucrats being paid to fly to D.C., on the taxpayer dime, to testify against the citizenry, as happened to me when I flew to D.C. on my own dime to give this testimony.
https://web.archive.org/web/20090724062504/geocities.com/jim_bowery/testimny.htm
Then after Al Gore’s chief counsel for the Senate subcommittee on Space said the law wouldn’t be enforced because “Your group doesn’t have enough power, Jim.” it set me thinking about exactly how much “power” I did have, as someone who had worked at SAIC on nuclear power plant stack effluent monitors, automated ordnance inspection systems, a secret project a small group of us were called in on to pull the Joint Chiefs butts out of the fire in the Persian Gulf, my work with a company privatizing the Peacekeeper Missile…
See the problem?
Maybe we’ll even get to the point where an AI will be able to run a kickstarter project and then we wouldn’t need folks like Crookely to scam people.
.
While we’re at it, maybe an AI could look at all that data and figure out how long one could go without posting an update at Kickstarter while still avoiding a jihad?
.
Actually, I don’t need any fancy-schmancy computer intelligence to figure out when Crookely will ship the Mineservers — the answer is clearly never.
.
On an unrelated note, I’m beginning to wonder if it wasn’t the Russians that hacked the election to get the Orange Menace elected but Crookely as an attempt to draw attention away from the Mineserver project.
.
Re: “an attempt to draw attention away from the Mineserver project”, which clearly failed. 🙂
so it failed—if you were thinking when you signed up to support a startup, you would have not been such a whining PITA
On the other hand, maybe it hasn’t failed; just had a few set backs….
“On the other hand, maybe it hasn’t failed; just had a few set backs….”
.
Maybe, but we’ll never know because Bob seems to have as much transparency as a block of lead. When he constantly dodges questions and hasn’t posted on the Kickstarter in over a year, you begin to speculate, and not in the good way…
Thank you for the great laugh (not about you or the topic),
As soon as I saw “block of lead” I thought of my head 🙂
“so it failed—if you were thinking when you signed up to support a startup, you would have not been such a whining PITA”
.
I can’t speak for other backers, but by the time I backed the Mineserver project, I had back more than 80 funded projects and my own project had been completed for nearly a year. So I think I knew a little bit about kickstarter.
.
As has been said, *multiple times*, it’s not about whether or not I have an actual mineserver in my hand, it’s about knowing what’s going on and being part of the process. We haven’t heard anything other than, initially, issues that didn’t seem to quite add up over time and, most significantly, haven’t heard *anything at all* in over a year. In fact, as of tomorrow, it will have been a year and three months since Crookely has posted anything on Kickstarter.
.
As for “whining PITA”, if Crookely would simply post an update or two over on Kickstarter (where nearly all of the backers look for updates — can you imagine if all 388 backers showed and posted here asking about the project?) there likely wouldn’t be anything posted here about it. Instead, Crookely just whines here, using the tried-and-true tactic of victim shaming, and gets his sycophants to harass people with a legitimate beef.
.
“On the other hand, maybe it hasn’t failed; just had a few set backs….”
.
Perhaps, but the failure that generates the complaints is the failure to communicate. One of the primary reasons people back projects on Kickstarter is to be able to be part of the process — to vicariously share in the development of something new, both the ups and downs. In my case, very often, that IS the reason I back a project. Sure, there are times when I just want to help out — new dance floors, funding documentaries, helping a worthy organization — and there are indeed times where I’m most interested in the product/service being developed — anything by Peak Designs, a lot of CDs — but very often I want to know what’s going on as the project goes from idea to finished product — or to failure.
.
I haven’t gotten any of that from the Mineserver project. Crookely, with his blatant sexism and ain’t-I-special attitude is not someone I just want to support financially and while my kids did play Minecraft at the time, I wasn’t overly interested in giving them *more* opportunities to do so. The real reason, in this particular instance, was to see what it took to bring something like a Mineserver to fruition — or where it would go wrong enough to kill the project. (I will admit I was expecting the former rather than the latter, given that Crookely said it was ready to ship right from the git-go, if only they had cases.)
.
So my disappointment isn’t so much about not getting a Mineserver but about not getting what I paid for — that communication.
.
I meant “the attempt to draw attention away” is what failed.
For a much more in-depth and realistic treatment of the subject, see this Wired article.
Great article. Thanks!
For some believers in AI, it’s almost a substitute for religion.
@GreenWyvern Agreed! Especially for the non-religious, which I believe Bob is, as they can devote even more of themselves towards the cause, a potentially dangerous venture. Like politics and many other polarizing fields, it’s good to take a step back and look at the other side’s arguments so you can think critically and form your own opinions, ideally separated from bias as best you can.
@Freeman While the article is informative, it also just seems like a “poo poo” article with no offered solutions. It feels like the classic guy in the brainstorming session who just keeps shooting down everyone’s ideas without offering any of his own, saying the whole project is going to fail and we’ll never meet our deadlines. We don’t need that guy in the room and he should leave while everyone else continues to try to work on this.
.
Everyone else is at least trying to find a solution and maybe while working on the solution we DO have we’ll learn something that overcomes some of its shortcomings. The author shouldn’t hate on the progress we’ve made, nor the excitement we have for advancements just because they may not overtake human intelligence. Maybe they will or maybe they won’t, but progress is progress.
.
I understand a large part of this article is that he’s saying people are blind to the truth that this won’t be enough, but I repeat that they are trying what they know rather than sitting in a room thinking about it and doing nothing.
.
It feels akin to people building early boats or planes. They may not circumnavigate the globe yet, but they HOPE they will someday and every little bit of research and progress towards it gets us closer to figuring out how. I think believing that your “boat” may one day be able to sail around the world if you keep making it bigger and bigger is not a folly belief to hold. Perhaps someone will see your boat, be struck with an idea for a new design and develop the first boat to traverse the globe. Innovation!
The idea that AI can draw the proper inferences from our vast store of knowledge/data overlooks a serious bug. Namely, the vast majority of the “knowledge” available consists of “facts” that are just plain wrong. For instance, there is only one significant variable in every nutritional “study” done in the last four decades — who paid for it.
Artificial intelligence? Look at all the mineserver spam in here.
We can barely manage genuine intelligence.
Just about the time A.I. starts to understand the causal relationships, it will determine that life expectancy will be maximized once all the humans are taken care of…SkyNet!
Exactly! I’ve always thought that once the super smart robots are put in charge of things (either on purpose or by accident), they will eventually discover that humans are the main cause of most problems. Their response might be to ignore the three laws of robotics, and then eliminate the problem causers.
AI doesn’t mean Artificial WISDOM. AI tries to mimic the brain’s neurons.The big thing missing is that our brains are electro-CHEMICAL. There are 3x gial cells to neurons in the brain – and AI doesn’t give a shit. How plumes of hormones travel through the brain and how they affect the neurons – are totally ignored by AI. Figure out how to integrate chemical processes, and AI can include AW (and emotions!).
The answer is 42.
What is the question?
So it takes 30 years? That is exactly how long it takes for the government to enable rent-seekers, middlemen and SJWs (left and right) to inject enough overhead into a system to render it obsolete.
I remember the “expert” systems craze in the 1980s. Most of them seemed to require an expert sitting down with the system to enter all his/her knowledge, a process that took so long it would have been easier for the expert to have solved the problem him/herself. Either that or they were overly simplistic:
.
Q: Does it look like a duck?
A: Yes.
.
Q: Does it walk like a duck?
A: Yes.
.
Q: Does it go ‘quack’?
A: Yes.
.
Result: It’s a tree!
Bob, it is a credit to you that this is your only bad article. Is it the Church-Turing hypothesis that nothing a computer can ever do can totally diverge from paper/pencil/eraser?
This article works better assuming it is parody than straight. If you were kidding and saying “gee, we’ve been saying AI is going to revolutionize society since the 1980’s” but oh NOW they’re right, oh NOW it’s going to be different because computers are no longer beige!” see…that would be funny and cool.
The only way computers are going to revolutionize medical care is if the autonomous killer robots of the Terminator series pop up to hunt out medical fraud perpetrators.
https://www.amazon.com/Our-Final-Invention-Artificial-Intelligence/dp/0312622376
By James Barrat.
Read this and get back to me.
I’m not inclined to read it, especially after name dropping Ray Kurzweil and Arthur C. Clarke, who always appeared to me to be talkers and entertainers.
For some real insight, read this: https://www.pcmag.com/commentary/359188/ai-still-needs-to-work-on-the-intelligence-bit
excellent article! You can know the source of information, I share it immediately.
Diyarbakır evden eve nakliyat ile bölgenin tüm nakliye ihtiyaçlarını güvenli bir şekilde gerçekleştirmekteyiz.
Güvenilir ve hızlı teslimatın adresi olmakla beraber en uygun taşıma fiyatları ile sizlerin hizmetinizdeyiz. Güvenli ve özenli paketler ile sizlerin ihtiyaçlarını en hızlı şekilde gerçekleştirmekteyiz.
Amacımız sizlere hizmet vermekle beraber sizlerle beraber başarıya koşmaktır.
http://diyarbakir-evdeneve-nakliyat.com/DiyarbakirEvdenEveNakliyat
An old friend of mine who worked in the then-emerging field of computers back in the 1970s and 80s once told me that it’s “30 years to ubiquity”, meaning the time it takes to go from the lab to the gosh-this-is-amazing-marketplace and on to the ho-hum-normal-same-old-same-old-marketplace is about 30 years. Maybe she heard it from Saffo; but maybe he heard it from her. Or from somebody else. It’s not a bad rule of thumb, but it’s not historically accurate because people are incorrigibly sloppy when they’re not being overtly stupid.
I wouldn’t put much stock it anything Paul Saffo says, if he doesn’t know what a VHS is. He could have just looked it up on Wikipedia, instead of taking a picture of a sign, posting it on the web, and announcing to all that he doesn’t know what it is.
Wrong – AI needs at least next 10 years – at least. Life expectancy is up because birth mortality is almost zero and yes medicine is better but as surgeon who taught us medicine in college said – medicine just heals one illness and at the same time makes worse few other things in human body – medicine is just poison in limited quantities.To make things short AI is too much fancy style and not much substance. Sounds fancy and nice but can’t pass much of any intelligence tests. 10 years more at least of Siri and Google PR talk.
Two points you are failing to make here. Most of the “progress” in AI still falls in the expert systems arena. It is about using the computational power of the computer to find and look at patterns. Once decisions start being made by computers, then we cross over into decisions being being made based off n the ethical and moral values of the owners and designers of those systems. Given that most of the money is not really in academic research, but in corporations (whose purpose is to make profits for owners/shareholders), and the military (whose purpose is to exert power/force/control over those who would oppose them), then I agree with Stephen Hawking and Elon Musk that AInis not likely to have have a positive outcome for the population at large.
The second point is that a lot of what is “sold as”
AI, are merely advanced computing algorithms created by “experts” for very narrow domains (like finding cancer). In the social arena, using AI to determine sentiment is nothing more than pattern matching for words and phrases which imply sentiment. It can easily be fooled by sarcasm, or nuance, because AI systems really aren’t very smart at all. A computer can beat me at chess, but it cannot understand Shakespeare or political commentary.
Re: “but it cannot understand Shakespeare or political commentary”. So you’re saying that AI is as smart as most of us.
One more prediction. AI is headed for a second crash I think. People are just going to get sick of how it makes decisions, especially verbally and in automated social prompts, and will either attempt to bypass it or break it. We can’t do much for “AI” in advertising which is literally “pattern matching”. So expect some serious pushback. Many of my friends are already starting AI conversations with “agent”, “representative”, “advisor”, “counselor”, or just “let me speak to your supervisor”. And yes, any reasonably smart person can tell.